Business object matching method and related equipment
By parsing the retrieved text and performing knowledge graph and visual feature vector retrieval in parallel, and integrating and rearranging candidate objects, the problem of business object matching deviation caused by the diversity of user input and incomplete information is solved, achieving higher accuracy and stability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- KINGDEE SMART TECH (SHENZHEN) CO LTD
- Filing Date
- 2026-02-04
- Publication Date
- 2026-05-15
AI Technical Summary
Existing technologies struggle to guarantee the accuracy and reliability of business processing results when faced with diverse user input formats, incomplete descriptive information, or high similarity of processing objects, thus impacting overall business processing efficiency.
By parsing the search text to generate search attribute information, and performing two search processes in parallel based on knowledge graphs and visual feature vectors, complementary candidate coverage is formed. Subsequently, the candidate business object set is fused and rearranged to determine the target business object.
It improves the accuracy and stability of business object matching results, enhances the reliability and efficiency of overall business processing, and reduces the risk of errors caused by single-path deviations.
Smart Images

Figure CN122045398A_ABST